Support Crosscut

Beneath downtown Seattle, Bertha faces biggest test

Bertha's about 20 percent there. Credit: WSDOT

The Seattle Mariners and the tunnel boring machine Bertha are following similar arcs lately. After years of wallowing in failure, both have shown a bit of a heartbeat in 2016.

Earlier this month, Bertha burrowed beneath the Alaskan Way viaduct without causing catastrophic damage. The elevated highway opened a week early and Bertha just kept on going. After a brief pause, the machine is moving again, drilling under downtown toward South Lake Union.

Several tunnel engineers say the progress is proof that we should put more trust in the machine’s operator, Seattle Tunnel Partners. “These machines are highly specialized,” says Dr. Reza Hedayat, an assistant professor at the Colorado School of Mines.

But a pattern of mechanical problems and operator errors have drained optimism in the machine, Seattle Tunnel Partners and the Washington Department of Transportation, which hired the company. As a result, we have dinner theater City Council meetings in which elected officials demand explanations about how they will safely burrow beneath downtown, only to have engineers respond, “It’s complicated.”

Many Seattleites simply wonder what will go wrong next.


A tunnel boring machine has two very important functions: to remove dirt and to prevent the dirt above from collapsing into the hole it’s just created. Few of us cared much about how Bertha did this until January, when the machine created enormous sinkholes near one of the most precarious major highways in the world. It turns out that the details are pretty important.

There are two types of tunnel boring machines, each of which deals with the whole collapsing earth thing a little differently. For the Seattle tunnel, WSDOT went with an Earth Pressure Balance Machine (EPBM). There is some disagreement surrounding whether that was the right choice, but we’ll get back to that.

To understand how an EPBM works, imagine pushing a couch up a flight of stairs: What’s keeping the pleather davenport from crashing down is the counteracting force of your body. Digging a tunnel project is like pushing a really heavy couch. The pressure in some cases can be many times normal atmospheric pressure.

To maintain the pressure, and prevent the earth in front of it from caving in, the machine does a bit of tunneling jujitsu. It fills a chamber behind the cutterhead with the dirt it’s excavating. So rather than earth getting siphoned out of the tunnel straight away, there’s a transition phase.

It’s a bit like an upright container in the bulk foods aisle, where you can control the amount of rice you want by opening the door more or less. Taking into account the speed of the machine and the pressure of the earth, tunnel operators can decide how wide to open that door — or in the case of Bertha, how quickly to spin a corkscrew conveyor, which carries the dirt out the back of the machine.

What’s so hard about that? Well, lots of things. First of all, balancing excavation with the forward progress of the machine is delicate. If not enough material is cleared out, the machine can cause the earth to heave upward. But excavate too quickly, and it could create voids in the soil and lead to a sinkhole. Open the flour container in the bulk food aisle all the way and you’ll create a little pocket of air running up through the whole container.

This, in fact, is exactly what happened on the waterfront in January. Seattle Tunnel Partners was apparently taking out too much soil, which led the massive hole to open in the ground above. This was not unforseen: several WSDOT employees predicted it days in advance when they saw how quickly STP was digging.

But there’s more to this than just managing the speed of the machine and the pressure behind the drill head. Staying in the bulk food aisle for a little longer, imagine that same container was segmented into different food products — rice, almonds, prunes — but still only had one door. If you open the hatch, the rice will come out much faster than the almonds or the prunes.

The ground underneath Seattle has about eight different food products — soils, in this case. The machine doesn’t have to deal with all eight simultaneously. But as it descends, which is what it’s doing right now, it will pass through different layers of soil. And because this machine is nearly six stories high — the biggest of its kind in the world — it may be churning through wildly varying kinds of soil all at once, as well as longer transition periods between the two. A 2006 paper out of the U.K. on sinkhole control points to these “interfaces” between soil types, saying it is “very difficult to maintain EPB pressure in these conditions.”

On top of the difficult soil conditions, tunnel machine operators must contend with the simple laws of physics. Swim six stories underwater and your eardrums could implode because of the pressure difference. It’s an apt comparison: Seattle’s soils are notoriously waterlogged, although water content varies, and to accurately make the calculate the water pressure, one needs to know how much water is actually in that soil.

So to return to my mangled food metaphor, tunneling under Seattle entails the equivalent of opening a bulk food container that contains up to eight different kinds of food, all of which behave differently, and maintaining the perfect, steady flow out the bottom. Meanwhile, the grocery clerk is sitting at the top, pushing down on the food with all his strength on one side, but hardly at all on the other.

The challenge is mitigated by essentially mixing all the soils into one, consistent muck — the equivalent of running the food through a blender before it goes into your bag, to beat this metaphor to death.


Impossible as all this may sound, tunnels are an important and common part of transportation infrastructure the world over. That’s to say that these are far from insurmountable problems. Sound Transit ran into similar problems when it bored beneath the Montlake Cut, but today we have a tunnel that will take you from downtown to the University of Washington in less time than in takes to 8 minutes. And finding the right earth pressure balance is not done on a napkin; unimaginably advanced algorithms do these calculations in real time.

“By EPB pressure standards, Seattle is nothing out of the ordinary,” says Doug Harding with the Robbins Company, which manufactures tunnel boring machines. The difference in pressure from the top to the bottom of the tunnel is enormous, to be sure, but Harding says it’s mostly a matter of plugging in the right calculations. “There are sensors at the top, middle and bottom,” he says. “The computers will tell the operator what the pressure should be.”

Harding agrees that the most difficult part of tunneling in Seattle is the soils. But Seattle Tunnel Partners has been consistent in its message that, as the machine descends, the soils will get easier to navigate.

Still, there are skeptics. Thom Neff, a Boston-based consultant with a doctorate in geotechnics, has become Seattle’s on-the-record tunnel doubter. In fact, he doubted its chances of success before it was cool to do so, whispering (or shouting) in the ear of then-Mayor Mike McGinn. So far, he’s looked pretty smart.

Despite recent successes and promises of better soils to come, Neff still doesn’t think Bertha is the machine to finish the job. “Primarily,” he says in an email, “because of the inability to accurately know the ground conditions in advance.”

A lot of the work of a tunnel is done before digging begins. Among the most important tasks is developing a strong understanding of the soil conditions. Without that, you don’t know what kind of bulk food you’re getting into. That work was done in Seattle, but Neff questions the accuracy. After all, WSDOT has already received a slap on the wrist for misclassifying a pipe in Bertha’s path (They thought it was one material when it turned out to be another, although no one has concluded whether that pipe contributed to the machine’s problems).

This is the point where trust in the parties is so essential. Was the analysis of soils done accurately? And will Bertha’s operators navigate them well?

After the most recent sinkhole, Washington Gov. Jay Inslee and WSDOT ordered Seattle Tunnel Partners to stop work until it swapped out its quality control team and several tunnel operators. That’s all well and good, except that several tunnel experts emphasized to Crosscut the importance of team members establishing good rapport with one another before proceeding under sensitive areas.

But beyond soil analysis and team competency, Neff argues that an EPBM was the wrong choice for this project. There is another machine called a liquid-supported mixshield that essentially adds another, partially filled tank behind the cutterhead operations. Goopy material and compressed air can be added or taken away from that chamber, which in theory creates a more adaptable pressure control system.

One of the losing bidders for the tunnel project, Martin Herrenknecht, has been highly critical of the decision to go with an EPBM, arguing that the machine will have trouble maintaining pressure in Seattle’s subterranean environs.

Harding, of the Robbins Company, dismisses the argument. He points to the issues with the Brightwater sewer tunneling project out near Woodinville as proof that mixshield machines can have difficulties, too. Operators there struggled mightily with sinkholes. STP representative have also stood by their choice.

For now, we seem to have little choice but to trust, and wait. At this writing, Bertha was drilling beneath the intersection of Western and Marion — right under the Post Office and a fancy furniture store. From there, it will head north toward Stewart before veering off toward South Lake Union. The plan for now is to finish by February, 2017.

That leaves an awful lot of time for things to go awry. And if so, who pays for the damage? And if the machine gets stuck, or breaks again? Then what?

For now, Bertha seems to be on a winning streak. If this continues, we can all forget about the difference between an EPBM and a mixshield, and the decision to choose one over the other. But if things go haywire, as they have many times already, rest assured that we’ll remember it for a long time.

Support Crosscut